Advances in cell-replacement therapy for Type I diabetes mellitus and a shortage of transplantable islets of Langerhans have focused interest on developing sources of insulin-producing cells, or β cells, appropriate for engraftment. One approach is the generation of functional β cells from pluripotent stem cells, such as, for example, embryonic stem cells.
In vertebrate embryonic development, a pluripotent cell gives rise to a group of cells comprising three germ layers (ectoderm, mesoderm, and endoderm) in a process known as gastrulation. Tissues such as, thyroid, thymus, pancreas, gut, and liver, will develop from the endoderm, via an intermediate stage. An intermediate stage in this process is the formation of definitive endoderm. Definitive endoderm cells express a number of markers, such as, HNF3beta, GATA4, MIXL1, CXCR4 and SOX17.
By the end of gastrulation, the endoderm is partitioned into anterior-posterior domains that can be recognized by the expression of a panel of factors that uniquely mark anterior, mid, and posterior regions of the endoderm. For example, Hhex, and Sox2 identify the anterior region while Cdx1, 2, and 4 identify the posterior half of the endoderm.
Migration of endoderm tissue brings the endoderm into close proximity with different mesodermal tissues that help in regionalization of the gut tube. This is accomplished by a plethora of secreted factors, such as FGFs, Wnts, TGF-Bs, retinoic acid (RA), and BMP ligands and their antagonists. For example, FGF4 and BMP promote Cdx2 expression in the presumptive hindgut endoderm and repress expression of the anterior genes Hhex and SOX2 (Development 2000, 127:1563-1567). WNT signaling has also been shown to work in parallel to FGF signaling to promote hindgut development and inhibit foregut fate (Development 2007, 134:2207-2217). Lastly, secreted retinoic acid by mesenchyme regulates the foregut-hindgut boundary (Curr Biol 2002, 12:1215-1220).
The level of expression of specific transcription factors may be used to designate the identity of a tissue. During transformation of the definitive endoderm into a primitive gut tube, the gut tube becomes regionalized into broad domains that can be observed at the molecular level by restricted gene expression patterns. For example, the regionalized pancreas domain in the gut tube shows a very high expression of PDX-1 and very low expression of CDX2 and SOX2. Similarly, the presence of high levels of Foxe1 are indicative of esophagus tissue; highly expressed in the lung tissue is NKX2.1; SOX2/Odd1 (OSR1) are highly expressed in stomach tissue; expression of PROX1/Hhex/AFP is high in liver tissue; SOX17 is highly expressed in biliary structure tissues; PDX1, NKX6.1/PTf1a, and NKX2.2 are highly expressed in pancreatic tissue; and expression of CDX2 is high in intestine tissue. The summary above is adapted from Dev Dyn 2009, 238:29-42 and Annu Rev Cell Dev Biol 2009, 25:221-251.
Formation of the pancreas arises from the differentiation of definitive endoderm into pancreatic endoderm (Annu Rev Cell Dev Biol 2009, 25:221-251; Dev Dyn 2009, 238:29-42). Dorsal and ventral pancreatic domains arise from the foregut epithelium. Foregut also gives rise to the esophagus, trachea, lungs, thyroid, stomach, liver, pancreas, and bile duct system.
Cells of the pancreatic endoderm express the pancreatic-duodenal homeobox gene PDX1. In the absence of PDX1, the pancreas fails to develop beyond the formation of ventral and dorsal buds. Thus, PDX1 expression marks a critical step in pancreatic organogenesis. The mature pancreas contains, among other cell types, exocrine tissue and endocrine tissue. Exocrine and endocrine tissues arise from the differentiation of pancreatic endoderm.
D'Amour et al. describes the production of enriched cultures of human embryonic stem cell-derived definitive endoderm in the presence of a high concentration of activin and low serum (Nature Biotechnol 2005, 23:1534-1541; U.S. Pat. No. 7,704,738). Transplanting these cells under the kidney capsule of mice resulted in differentiation into more mature cells with characteristics of endodermal tissue (U.S. Pat. No. 7,704,738). Human embryonic stem cell-derived definitive endoderm cells can be further differentiated into PDX1 positive cells after addition of FGF-10 and retinoic acid (U.S. Patent Publication No. 2005/0266554A1). Subsequent transplantation of these pancreatic precursor cells under the kidney capsule of immune deficient mice resulted in formation of functional pancreatic endocrine cells following a 3-4 month maturation phase (U.S. Pat. No. 7,993,920 and U.S. Pat. No. 7,534,608).
Fisk et al. report a system for producing pancreatic islet cells from human embryonic stem cells (U.S. Pat. No. 7,033,831). In this case, the differentiation pathway was divided into three stages. Human embryonic stem cells were first differentiated to endoderm using a combination of sodium butyrate and activin A (U.S. Pat. No. 7,326,572). The cells were then cultured with BMP antagonists, such as Noggin, in combination with EGF or betacellulin to generate PDX1 positive cells. The terminal differentiation was induced by nicotinamide.
Small molecule inhibitors have also been used for induction of pancreatic endocrine precursor cells. For example, small molecule inhibitors of TGF-B receptor and BMP receptors (Development 2011, 138:861-871; Diabetes 2011, 60:239-247) have been used to significantly enhance number of pancreatic endocrine cells. In addition, small molecule activators have also been used to generate definitive endoderm cells or pancreatic precursor cells (Curr Opin Cell Biol 2009, 21:727-732; Nature Chem Biol 2009, 5:258-265).
Previous attempts at the induction of pancreatic precursor cells from human embryonic stem cells have highlighted the importance of co-expression of PDX-1 and NKX6.1 in correctly identifying pancreatic endoderm. However, while the art has identified the population of cells positive in expression of PDX-1 and NKX6.1 to be low or absent of CDX2 expression, previous reports have failed to test for presence of markers just anterior to the developing pancreas. SOX2, which marks the anterior endoderm, is not expressed in adult islets and is expressed at a very low level in developing pancreas (Diabetes 2005, 54:3402-4309). In contrast, some of the examples in this application disclose cell populations where at least 30% of the pancreatic endoderm cells generated from human embryonic stem cells are positive for the expression of PDX-1 and NKX6.1, and negative for the expression of CDX2 and SOX2.
All of the previous attempts to generate functional pancreatic beta cells have fallen short of attaining cells with characteristics of mature beta cells. Hallmarks of mature beta cells include expression of single hormonal insulin, correct processing of proinsulin into insulin and C-peptide, strong expression of PDX-1 and NKX6.1, appropriate insulin release in response to glucose, expression of glucose transporters, and high expression of glucokinase. All of the previous reports have resulted in endocrine cells that produce two or more of the pancreatic hormones. For example, D'Amour et al (Nature Biotech 2006, 24:1392-1401) report the generation of a cell population comprising ˜10% insulin positive cells and ˜20% endocrine cells as measured by synaptophysin. Similar reports by others (Cell Res 2009, 19:429-438; Stem Cells 2007, 25:1940-1953; Diabetes Obes Metab 2008, 10:186-194) have also shown differentiation of pluripotent cells to non-functional insulin positive cells. Indeed, recent studies have clearly established that transplantation of polyhormonal cells in Severe Combined ImmunoDeficiency (SCID) mice did not result in generation of functional beta cells (Diabetes 2011, 60:239-247; Nature Biotech 2011, 29:750-756). While in human fetal pancreas a fraction (˜10-20%) of endocrine cells are polyhormonal cells; polyhormonal cells disappear in adult human pancreas (Histochem Cell Biol 1999, 112:147-153; J Histochem Cytochem 2009, 57:811-824).
As the burgeoning field of regenerative medicine continues to mature, a method for the formation of terminally differentiated, appropriately regulated pancreatic endocrine cells is highly desirable. We demonstrate here that with appropriate and defined manipulation of culture conditions, and precise timing of the addition of activators/inhibitors of various pathways, human embryonic stem cells can be differentiated in vitro into functional pancreatic beta cells. In particular, precise timing of BMP inhibition, using a gradient of retinoic acid along with the use of Vitamin C proved effective in generation of single hormonal pancreatic endocrine cells.